Serveur d'exploration sur l'OCR

Attention, ce site est en cours de développement !
Attention, site généré par des moyens informatiques à partir de corpus bruts.
Les informations ne sont donc pas validées.

Off-line cursive handwriting recognition using hidden Markov models

Identifieur interne : 000A45 ( PascalFrancis/Corpus ); précédent : 000A44; suivant : 000A46

Off-line cursive handwriting recognition using hidden Markov models

Auteurs : H. Bunke ; M. Roth ; E. G. Schukat-Talamazzini

Source :

RBID : Pascal:95-0540564

Descripteurs français

English descriptors

Abstract

A method for the off-line recognition of cursive handwriting based on hidden Markov models (HMMs) is described. The features used in the HMMs are based on the arcs of skeleton graphs of the words to be recognized. An algorithm is applied to the skeleton graph of a word that extracts the edges in a particular order. Given the sequence of edges extracted from the skeleton graph, each edge is transformed into a 10-dimensional feature vector. The features represent information about the location of an edge relative to the four reference lines, its curvature and the degree of the nodes incident to the considered edge. The linear model was adopted as basic HMM topology. Each letter of the alphabet is represented by a linear HMM. Given a dictionary of fixed size, an HMM for each dictionary word is built by sequential concatenation of the HMMs representing the individual letters of the word. Training of the HMMs is done by means of the Baum-Welch algorithm, while the Viterbi algorithm is used for recognition. An average correct recognition rate of over 98% on the word level has been achieved in experiments with cooperative writers using two dictionaries of 150 words each.

Notice en format standard (ISO 2709)

Pour connaître la documentation sur le format Inist Standard.

pA  
A01 01  1    @0 0031-3203
A02 01      @0 PTNRA8
A03   1    @0 Pattern recogn.
A05       @2 28
A06       @2 9
A08 01  1  ENG  @1 Off-line cursive handwriting recognition using hidden Markov models
A11 01  1    @1 BUNKE (H.)
A11 02  1    @1 ROTH (M.)
A11 03  1    @1 SCHUKAT-TALAMAZZINI (E. G.)
A14 01      @1 Univ. Bern, Inst. Informatik angewandte Mathematik @2 3012 Bern @3 CHE @Z 1 aut. @Z 2 aut.
A20       @1 1399-1413
A21       @1 1995
A23 01      @0 ENG
A43 01      @1 INIST @2 15220 @5 354000054444100070
A44       @0 0000
A45       @0 40 ref.
A47 01  1    @0 95-0540564
A60       @1 P
A61       @0 A
A64 01  1    @0 Pattern recognition
A66 01      @0 GBR
C01 01    ENG  @0 A method for the off-line recognition of cursive handwriting based on hidden Markov models (HMMs) is described. The features used in the HMMs are based on the arcs of skeleton graphs of the words to be recognized. An algorithm is applied to the skeleton graph of a word that extracts the edges in a particular order. Given the sequence of edges extracted from the skeleton graph, each edge is transformed into a 10-dimensional feature vector. The features represent information about the location of an edge relative to the four reference lines, its curvature and the degree of the nodes incident to the considered edge. The linear model was adopted as basic HMM topology. Each letter of the alphabet is represented by a linear HMM. Given a dictionary of fixed size, an HMM for each dictionary word is built by sequential concatenation of the HMMs representing the individual letters of the word. Training of the HMMs is done by means of the Baum-Welch algorithm, while the Viterbi algorithm is used for recognition. An average correct recognition rate of over 98% on the word level has been achieved in experiments with cooperative writers using two dictionaries of 150 words each.
C02 01  1    @0 001D02C03
C03 01  X  FRE  @0 Reconnaissance caractère @5 01
C03 01  X  ENG  @0 Character recognition @5 01
C03 01  X  SPA  @0 Reconocimiento carácter @5 01
C03 02  X  FRE  @0 Caractère manuscrit @5 02
C03 02  X  ENG  @0 Manuscript character @5 02
C03 02  X  SPA  @0 Carácter manuscrito @5 02
C03 03  X  FRE  @0 Alphabet @5 03
C03 03  X  ENG  @0 Alphabet @5 03
C03 03  X  SPA  @0 Alfabeto @5 03
C03 04  X  FRE  @0 Modèle Markov @5 04
C03 04  X  ENG  @0 Markov model @5 04
C03 04  X  SPA  @0 Modelo Markov @5 04
C03 05  X  FRE  @0 Cursive script recognition @4 INC @5 72
C03 06  X  FRE  @0 Off line recognition @4 INC @5 73
C03 07  X  FRE  @0 Hidden Markov model @4 INC @5 74
C03 08  X  FRE  @0 Skeleton graph @4 INC @5 75
C03 09  X  FRE  @0 OCR @4 CD @5 96
C03 09  X  ENG  @0 OCR @4 CD @5 96
N21       @1 310

Format Inist (serveur)

NO : PASCAL 95-0540564 INIST
ET : Off-line cursive handwriting recognition using hidden Markov models
AU : BUNKE (H.); ROTH (M.); SCHUKAT-TALAMAZZINI (E. G.)
AF : Univ. Bern, Inst. Informatik angewandte Mathematik/3012 Bern/Suisse (1 aut., 2 aut.)
DT : Publication en série; Niveau analytique
SO : Pattern recognition; ISSN 0031-3203; Coden PTNRA8; Royaume-Uni; Da. 1995; Vol. 28; No. 9; Pp. 1399-1413; Bibl. 40 ref.
LA : Anglais
EA : A method for the off-line recognition of cursive handwriting based on hidden Markov models (HMMs) is described. The features used in the HMMs are based on the arcs of skeleton graphs of the words to be recognized. An algorithm is applied to the skeleton graph of a word that extracts the edges in a particular order. Given the sequence of edges extracted from the skeleton graph, each edge is transformed into a 10-dimensional feature vector. The features represent information about the location of an edge relative to the four reference lines, its curvature and the degree of the nodes incident to the considered edge. The linear model was adopted as basic HMM topology. Each letter of the alphabet is represented by a linear HMM. Given a dictionary of fixed size, an HMM for each dictionary word is built by sequential concatenation of the HMMs representing the individual letters of the word. Training of the HMMs is done by means of the Baum-Welch algorithm, while the Viterbi algorithm is used for recognition. An average correct recognition rate of over 98% on the word level has been achieved in experiments with cooperative writers using two dictionaries of 150 words each.
CC : 001D02C03
FD : Reconnaissance caractère; Caractère manuscrit; Alphabet; Modèle Markov; Cursive script recognition; Off line recognition; Hidden Markov model; Skeleton graph; OCR
ED : Character recognition; Manuscript character; Alphabet; Markov model; OCR
SD : Reconocimiento carácter; Carácter manuscrito; Alfabeto; Modelo Markov
LO : INIST-15220.354000054444100070
ID : 95-0540564

Links to Exploration step

Pascal:95-0540564

Le document en format XML

<record>
<TEI>
<teiHeader>
<fileDesc>
<titleStmt>
<title xml:lang="en" level="a">Off-line cursive handwriting recognition using hidden Markov models</title>
<author>
<name sortKey="Bunke, H" sort="Bunke, H" uniqKey="Bunke H" first="H." last="Bunke">H. Bunke</name>
<affiliation>
<inist:fA14 i1="01">
<s1>Univ. Bern, Inst. Informatik angewandte Mathematik</s1>
<s2>3012 Bern</s2>
<s3>CHE</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author>
<name sortKey="Roth, M" sort="Roth, M" uniqKey="Roth M" first="M." last="Roth">M. Roth</name>
<affiliation>
<inist:fA14 i1="01">
<s1>Univ. Bern, Inst. Informatik angewandte Mathematik</s1>
<s2>3012 Bern</s2>
<s3>CHE</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author>
<name sortKey="Schukat Talamazzini, E G" sort="Schukat Talamazzini, E G" uniqKey="Schukat Talamazzini E" first="E. G." last="Schukat-Talamazzini">E. G. Schukat-Talamazzini</name>
</author>
</titleStmt>
<publicationStmt>
<idno type="wicri:source">INIST</idno>
<idno type="inist">95-0540564</idno>
<date when="1995">1995</date>
<idno type="stanalyst">PASCAL 95-0540564 INIST</idno>
<idno type="RBID">Pascal:95-0540564</idno>
<idno type="wicri:Area/PascalFrancis/Corpus">000A45</idno>
</publicationStmt>
<sourceDesc>
<biblStruct>
<analytic>
<title xml:lang="en" level="a">Off-line cursive handwriting recognition using hidden Markov models</title>
<author>
<name sortKey="Bunke, H" sort="Bunke, H" uniqKey="Bunke H" first="H." last="Bunke">H. Bunke</name>
<affiliation>
<inist:fA14 i1="01">
<s1>Univ. Bern, Inst. Informatik angewandte Mathematik</s1>
<s2>3012 Bern</s2>
<s3>CHE</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author>
<name sortKey="Roth, M" sort="Roth, M" uniqKey="Roth M" first="M." last="Roth">M. Roth</name>
<affiliation>
<inist:fA14 i1="01">
<s1>Univ. Bern, Inst. Informatik angewandte Mathematik</s1>
<s2>3012 Bern</s2>
<s3>CHE</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author>
<name sortKey="Schukat Talamazzini, E G" sort="Schukat Talamazzini, E G" uniqKey="Schukat Talamazzini E" first="E. G." last="Schukat-Talamazzini">E. G. Schukat-Talamazzini</name>
</author>
</analytic>
<series>
<title level="j" type="main">Pattern recognition</title>
<title level="j" type="abbreviated">Pattern recogn.</title>
<idno type="ISSN">0031-3203</idno>
<imprint>
<date when="1995">1995</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
<seriesStmt>
<title level="j" type="main">Pattern recognition</title>
<title level="j" type="abbreviated">Pattern recogn.</title>
<idno type="ISSN">0031-3203</idno>
</seriesStmt>
</fileDesc>
<profileDesc>
<textClass>
<keywords scheme="KwdEn" xml:lang="en">
<term>Alphabet</term>
<term>Character recognition</term>
<term>Manuscript character</term>
<term>Markov model</term>
<term>OCR</term>
</keywords>
<keywords scheme="Pascal" xml:lang="fr">
<term>Reconnaissance caractère</term>
<term>Caractère manuscrit</term>
<term>Alphabet</term>
<term>Modèle Markov</term>
<term>Cursive script recognition</term>
<term>Off line recognition</term>
<term>Hidden Markov model</term>
<term>Skeleton graph</term>
<term>OCR</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front>
<div type="abstract" xml:lang="en">A method for the off-line recognition of cursive handwriting based on hidden Markov models (HMMs) is described. The features used in the HMMs are based on the arcs of skeleton graphs of the words to be recognized. An algorithm is applied to the skeleton graph of a word that extracts the edges in a particular order. Given the sequence of edges extracted from the skeleton graph, each edge is transformed into a 10-dimensional feature vector. The features represent information about the location of an edge relative to the four reference lines, its curvature and the degree of the nodes incident to the considered edge. The linear model was adopted as basic HMM topology. Each letter of the alphabet is represented by a linear HMM. Given a dictionary of fixed size, an HMM for each dictionary word is built by sequential concatenation of the HMMs representing the individual letters of the word. Training of the HMMs is done by means of the Baum-Welch algorithm, while the Viterbi algorithm is used for recognition. An average correct recognition rate of over 98% on the word level has been achieved in experiments with cooperative writers using two dictionaries of 150 words each.</div>
</front>
</TEI>
<inist>
<standard h6="B">
<pA>
<fA01 i1="01" i2="1">
<s0>0031-3203</s0>
</fA01>
<fA02 i1="01">
<s0>PTNRA8</s0>
</fA02>
<fA03 i2="1">
<s0>Pattern recogn.</s0>
</fA03>
<fA05>
<s2>28</s2>
</fA05>
<fA06>
<s2>9</s2>
</fA06>
<fA08 i1="01" i2="1" l="ENG">
<s1>Off-line cursive handwriting recognition using hidden Markov models</s1>
</fA08>
<fA11 i1="01" i2="1">
<s1>BUNKE (H.)</s1>
</fA11>
<fA11 i1="02" i2="1">
<s1>ROTH (M.)</s1>
</fA11>
<fA11 i1="03" i2="1">
<s1>SCHUKAT-TALAMAZZINI (E. G.)</s1>
</fA11>
<fA14 i1="01">
<s1>Univ. Bern, Inst. Informatik angewandte Mathematik</s1>
<s2>3012 Bern</s2>
<s3>CHE</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</fA14>
<fA20>
<s1>1399-1413</s1>
</fA20>
<fA21>
<s1>1995</s1>
</fA21>
<fA23 i1="01">
<s0>ENG</s0>
</fA23>
<fA43 i1="01">
<s1>INIST</s1>
<s2>15220</s2>
<s5>354000054444100070</s5>
</fA43>
<fA44>
<s0>0000</s0>
</fA44>
<fA45>
<s0>40 ref.</s0>
</fA45>
<fA47 i1="01" i2="1">
<s0>95-0540564</s0>
</fA47>
<fA60>
<s1>P</s1>
</fA60>
<fA61>
<s0>A</s0>
</fA61>
<fA64 i1="01" i2="1">
<s0>Pattern recognition</s0>
</fA64>
<fA66 i1="01">
<s0>GBR</s0>
</fA66>
<fC01 i1="01" l="ENG">
<s0>A method for the off-line recognition of cursive handwriting based on hidden Markov models (HMMs) is described. The features used in the HMMs are based on the arcs of skeleton graphs of the words to be recognized. An algorithm is applied to the skeleton graph of a word that extracts the edges in a particular order. Given the sequence of edges extracted from the skeleton graph, each edge is transformed into a 10-dimensional feature vector. The features represent information about the location of an edge relative to the four reference lines, its curvature and the degree of the nodes incident to the considered edge. The linear model was adopted as basic HMM topology. Each letter of the alphabet is represented by a linear HMM. Given a dictionary of fixed size, an HMM for each dictionary word is built by sequential concatenation of the HMMs representing the individual letters of the word. Training of the HMMs is done by means of the Baum-Welch algorithm, while the Viterbi algorithm is used for recognition. An average correct recognition rate of over 98% on the word level has been achieved in experiments with cooperative writers using two dictionaries of 150 words each.</s0>
</fC01>
<fC02 i1="01" i2="1">
<s0>001D02C03</s0>
</fC02>
<fC03 i1="01" i2="X" l="FRE">
<s0>Reconnaissance caractère</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="ENG">
<s0>Character recognition</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="SPA">
<s0>Reconocimiento carácter</s0>
<s5>01</s5>
</fC03>
<fC03 i1="02" i2="X" l="FRE">
<s0>Caractère manuscrit</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="ENG">
<s0>Manuscript character</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="SPA">
<s0>Carácter manuscrito</s0>
<s5>02</s5>
</fC03>
<fC03 i1="03" i2="X" l="FRE">
<s0>Alphabet</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="ENG">
<s0>Alphabet</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="SPA">
<s0>Alfabeto</s0>
<s5>03</s5>
</fC03>
<fC03 i1="04" i2="X" l="FRE">
<s0>Modèle Markov</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="ENG">
<s0>Markov model</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="SPA">
<s0>Modelo Markov</s0>
<s5>04</s5>
</fC03>
<fC03 i1="05" i2="X" l="FRE">
<s0>Cursive script recognition</s0>
<s4>INC</s4>
<s5>72</s5>
</fC03>
<fC03 i1="06" i2="X" l="FRE">
<s0>Off line recognition</s0>
<s4>INC</s4>
<s5>73</s5>
</fC03>
<fC03 i1="07" i2="X" l="FRE">
<s0>Hidden Markov model</s0>
<s4>INC</s4>
<s5>74</s5>
</fC03>
<fC03 i1="08" i2="X" l="FRE">
<s0>Skeleton graph</s0>
<s4>INC</s4>
<s5>75</s5>
</fC03>
<fC03 i1="09" i2="X" l="FRE">
<s0>OCR</s0>
<s4>CD</s4>
<s5>96</s5>
</fC03>
<fC03 i1="09" i2="X" l="ENG">
<s0>OCR</s0>
<s4>CD</s4>
<s5>96</s5>
</fC03>
<fN21>
<s1>310</s1>
</fN21>
</pA>
</standard>
<server>
<NO>PASCAL 95-0540564 INIST</NO>
<ET>Off-line cursive handwriting recognition using hidden Markov models</ET>
<AU>BUNKE (H.); ROTH (M.); SCHUKAT-TALAMAZZINI (E. G.)</AU>
<AF>Univ. Bern, Inst. Informatik angewandte Mathematik/3012 Bern/Suisse (1 aut., 2 aut.)</AF>
<DT>Publication en série; Niveau analytique</DT>
<SO>Pattern recognition; ISSN 0031-3203; Coden PTNRA8; Royaume-Uni; Da. 1995; Vol. 28; No. 9; Pp. 1399-1413; Bibl. 40 ref.</SO>
<LA>Anglais</LA>
<EA>A method for the off-line recognition of cursive handwriting based on hidden Markov models (HMMs) is described. The features used in the HMMs are based on the arcs of skeleton graphs of the words to be recognized. An algorithm is applied to the skeleton graph of a word that extracts the edges in a particular order. Given the sequence of edges extracted from the skeleton graph, each edge is transformed into a 10-dimensional feature vector. The features represent information about the location of an edge relative to the four reference lines, its curvature and the degree of the nodes incident to the considered edge. The linear model was adopted as basic HMM topology. Each letter of the alphabet is represented by a linear HMM. Given a dictionary of fixed size, an HMM for each dictionary word is built by sequential concatenation of the HMMs representing the individual letters of the word. Training of the HMMs is done by means of the Baum-Welch algorithm, while the Viterbi algorithm is used for recognition. An average correct recognition rate of over 98% on the word level has been achieved in experiments with cooperative writers using two dictionaries of 150 words each.</EA>
<CC>001D02C03</CC>
<FD>Reconnaissance caractère; Caractère manuscrit; Alphabet; Modèle Markov; Cursive script recognition; Off line recognition; Hidden Markov model; Skeleton graph; OCR</FD>
<ED>Character recognition; Manuscript character; Alphabet; Markov model; OCR</ED>
<SD>Reconocimiento carácter; Carácter manuscrito; Alfabeto; Modelo Markov</SD>
<LO>INIST-15220.354000054444100070</LO>
<ID>95-0540564</ID>
</server>
</inist>
</record>

Pour manipuler ce document sous Unix (Dilib)

EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/PascalFrancis/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000A45 | SxmlIndent | more

Ou

HfdSelect -h $EXPLOR_AREA/Data/PascalFrancis/Corpus/biblio.hfd -nk 000A45 | SxmlIndent | more

Pour mettre un lien sur cette page dans le réseau Wicri

{{Explor lien
   |wiki=    Ticri/CIDE
   |area=    OcrV1
   |flux=    PascalFrancis
   |étape=   Corpus
   |type=    RBID
   |clé=     Pascal:95-0540564
   |texte=   Off-line cursive handwriting recognition using hidden Markov models
}}

Wicri

This area was generated with Dilib version V0.6.32.
Data generation: Sat Nov 11 16:53:45 2017. Site generation: Mon Mar 11 23:15:16 2024